Crystal structure of ammonium hexafluorocerate (IV),(NH4) 2CeF6

Robert Reynolds Ryan, Allen C. Larson, and F. H. Kruse. Inorg. ... Tiffany A. Pham , Alison B. Altman , S. Chantal E. Stieber , Corwin H. Booth , Stos...
0 downloads 0 Views 433KB Size
AMMONIUM HEXAFLUOKOCERATE(IV) 33

Vol. 8, No. 1, January 1969 solvated Yb(CSH702)3(H20) structure,3 and confidence in the reality of such features of the polyhedron as the deviations of the ytterbium atom from the mean planes of the ligands may be obtained by comparing the two structures. The benzene molecule is centered on the center of symmetry a t 0, l/2, l/2. The mean deviation of the carbon atoms from the mean plane of the benzene ring is 0.019 A, and the sum of the interior angles of the ring is 719.6'. Discrepancies in the carbon-carbon bond lengths (Table 111) of several standard deviations must be interpreted as indicative of systematic errors in the data, such as uncorrected adsorption, rather

than as implying any actual distortion of the molecule. The shortest distance between a benzene carbon atom and another carbon or oxygen atom is 3.53 if for the CB(1)-O(4) contact. The volume occupied by two tris(acety1acetonato)aquoytterbium(II1) molecules in the unsolvated structure3 is 931.8 A3. The molecular volume of benzene, based on the density of the liquid a t 25', is 147.6 A3. The sum of these volumes is 1079.4 k3,as compared with a unit cell volume of 1077.9A3in the crystal structure of the benzene solvate. These figures suggest that the efficiency of packing of the molecules is about the same for the solvated and unsolvated structures.

CONTRIBUTION FROM THE Los ALAMOS SCIENTIFIC LABORATORY, OB CALIFORNIA, Los ALAMOS, NEW MEXICO 87544 UNIVERSITY

The Crystal Structure of Ammonium Hexafluorocerate(IV), (NH,),CeF,l BY ROBERT R. RYAN, ALLEN C. LARSON,

AND

F. H. KRUSE

Received June 5, 1968 The crystal structure of ammonium hexafluorocerate(IV), (NH&CeFa, has been determined from counter measurements of 561 reflections and refined by full-matrix least squares to an R factor of 0.032. The space group is Pbcn (no. 60); D, = 3.05 and D, = 3.05 ==! 0.02 g/cm3. The unit cell has dimensions a = 7.026 (3),b = 12.098 (6), and c = 7.439 (8) A; Z = 4. Of the 561 reflections measured, 389 were judged to be observed. The structure consists of infinite chains of eightfold coordinated cerium atoms parallel to the c axis, joined by NH4+ ions. The fluorine configuration about the Ce can be roughly described as a square antiprism. The average Ce-F distance is 2.26 -k and the average F-F distance for nearest neighbors within an antiprism is 2.77 -k. Hydrogen bonding causes this particular structure to be stabilized over the trigonal-dodecahedral arrangement found in the similar RbaUFB and KzZrFe compounds.

Introduction The preparation of ammonium hexafluorocerate(1V) was first reported by Besse and Capestaq2 who indexed the powder pattern on the basis of a cubic cell with a' = 10.52 k and claimed a measured density of 3.27 g/cm3. For 2 = 8, D, = 3.31 g/cm3 for this cell. Penneman3 of this laboratory noticed that these values for the volume and 2 give a very low value for the fluorine volume, as compared to similar actinide fluoride compounds. Further, there is no evidence for cubic phases of any other 2 : l NH4F-actinide(IV) fluoride complexes. He was therefore prompted to prepare this material. Optical examination showed single crystals of (NH4)d2eF6to be biaxial and thus this compound must have orthorhombic or lower symmetry. Subsequent examination of precession films showed the symmetry to be orthorhombic with cell dimensions which are nearly those which would permit indexing of powder data on the basis of the cubic cell of Besse and Capestan,2 ;.e., a'/a = 3/2, a'/b = d 3 / 2 , and a'/c = ds. The single-crystal X-ray investigation of this compound was undertaken for three reasons: (1) No (1) This work was sponsored by the U. S. Atomic Energy Commission. (2) J. P. Besse and M. Capestan, Bull. SOC.Chim.France, 577 (1966). (3) R. A. Penneman, unpublished results, t o be submitted for publication.

compound of known structure could be found with which it is isostructural. ( 2 ) Comparison of the powder patterns indicated it is isostructural with the a phase of (NH4)zUFs for which no single crystals were then available. (Subsequent comparisons of singlecrystal data, when they became available, confirmed that the two compounds were indeed isostructural. This is of interest because of the existence of four closely related polymorphs for (NH4)2UFeover a narrow temperature range4 (-70'), and the elucidation of these structures should provide information concerning the energy relationships between some of the possible coordination polyhedra for uranium.) (3) It is also of interest that after extensive investigation the only phase to be found for the Ce compound is the a phase. It is hoped that the knowledge of this structure plus the structure of the y-(NH&UFe (room-temperature stable phase) compound currently being worked on will provide some information concerning these questions. Experimental Section Crystals were prepared by cooling a saturated solution of (NH&CeFo in a 15% NHIF-H~Osolution. A needlelike crystal was selected and cut to a length suitable for examination by X(4) R. A. Penneman, F. H.Kruse, R. S. George, and J. S. Coleman, Inorg. Chcm., 8,309 (1964).

34 ROBERT R. RYAN,ALLEN C. LARSON, AND F. H. KRUSE ray methods. The resulting 0.19 X 0.11 X 0 .07 mm crystal was mounted on the a axis and was coated with Fluorolube to arrest an apparent slow decomposition which took place upon contact of the material with air. The faces whose normals were perpendicular to the needle direction were of the prism (011 Precession photographs taken with Mo Ka:radiation showed orthorhombic symmetry with the following systematic extinctions: OkZ reflections, k odd; h0Z reflections, I odd; hkO reflections, h k odd. The space group is thus uniquely Pbcn (no. BO), D2lIL4. The cell constants a t room temperature were obtained by least-squares calculations from 13 high-order reflections measured with a carefully aligned single-crystal orienter on a General Electroc Co. XRD-5 diffraction unit using Mo Kml radiation JX(MoKa1)0.70930 A ] . The cell constants are a = 7.026 ( 3 ) , b = 12.098 (61, and c = 7.439 (8) d, where the figure in parentheses is the least-squares standard error of the least significant digit. The calculated density with 2 = 4 is 3.05 g/cm3; the density measured by flotation methods in mixtures of diiodomethane and dibromomethane was 3.05 i 0.02 g/cm3. Four molecules per unit cell for this space group demand twofold or inversion symmetry for the cerium atom position. w scans through several reflections showed the crystal mosaicity was sufficiently low. Intensities of reflections with 28 5 45” for hkl planes were measured using Mo Ka:radiation, Zr-Y balanced filters, takeoff angle of 5 ” , the stationary-crystal, stationary-counter mode, and a counting time of 20 sec with each filter. 4 few reflections from hkj planes were measured in order to have data to check the absorption correction. The counting rates were such that attenuation of the primary beam was unnecessary. Because of decomposition of the sample, the tube current was adjusted during the course of the data collection to keep the count for a standard reflection within 1% of its starting value. Without this adjustment the intensity of the standard reflection would have decreased by approximately 10% during the course of the data collection process. The crystal was checked periodically for misalignment and centering. Of the 561 reflections measured, 389 were observed according to thc criterion I > 3 ~ 4 1 ) . The standard deviation, o(l),of the corrected intensity, I , was estimated from

1.

+

u(1) =

[T

+ B + B’]”2 +

where T is the total count, I = T - ( B B ’ ) , B is the background as measured a t the reflection with the Y filter, and B’ is the background taken from an empirical 20-dependent background curve. The empirical curve was determined by averaging the background a t a given 28 value and several values of x and @ for which the Bragg condition was not satisfied. The absorption correction was made by the Busing and Levy5 method using Burnham’s6 program modified for single-crystal orienter geometry.? The linear absorption coefficient is 73.2 cm-l and the calculated transmission factors vary from 0.561 to 0.626.

Solution and Refinement of the Structure The position of the cerium atoms was determined from the Patterson map to be 0, 0.04, ‘/4 and was refined by least-squares methods to be 0, 0.0414, l / 4 . The structure is C-centered if only the cerium atoms are considered so that a Fourier map phased on the cerium showed a false twofold rotation axis a t y = l / z , z = 1/2. Since there were four atoms whose positions were unknown, there were Z 4 ways to choose the structure from the difference Fourier map. Although some of the models were slightly less desirable from a chemical point of view, the proximity of the peaks was such that none of the possible structures could be rejected on the ( 5 ) W. R. Busing and H. A. Levy, Acta C ~ y s l . 10, , 180 (1557). (6) C. W. Burnham, IUC World List of Crystallographic Computer I’iugi-ams, Program 338, 1562. (7) A. C. Larson, D. T. Cromer, and R. B. Roof, Jr., Report LA-3043, Los Alamos Scientific Laboratory, 1964.

Inorganic Chemistry

basis of chemical evidence alone. Phasing the difference map on the C-centered cerium atom leaves one degree of freedom in fixing the origin so that the choice between two of the peaks is arbitrary. This arbitrary choice was made for F(3)and the remaining eight models were refined using anisotropic temperature factors on the cerium atom only. The nitrogen atom always refined to the same position no matter which of the two possible peaks was chosen as the starting position. Of the remaining four models, three had a conventional unweighted R factor

of 0.060 and for the remaining model the li factor was

0.035. The corresponding weighted

R‘ factors

mere 0.050 and 0.033. The difference in R factors plus statistical tests on the ratio of the square roots of the weighted residuals8 dictated the choice of the correct model. The full-matrix least-squares refinement of all nonhydrogen positions and temperature factors was car~ ried out using a newly developed p r ~ g r a m . The quantity minimized was Zw,(Fo - K F J i 2 ,where w 6 is ~ / C J ~ ( F . )Scattering .~~ factors given in the “International Critical Tables for X-Ray Crystallography,” 1962, were used for fluorine and nitrogen atoms. Cromer and Waber’sl’ scattering factors and Cromer’s12 anomalous dispersion terms mere used for cerium. The final R with all atoms having isotropic factors was 0.059 (0.051), the R with only the cerium atom anisotropic was 0.035 (0.032), and the R with all atoms with anisotropic temperature factors was 0.032 (0.030), where the values in parentheses are the weighted R‘ factors. Hamilton’ss significance tests reject the hypothesis that all atoms are isotropic a t the 0.005 significance level while the hypothesis that only the cerium atoms are anisotropic can be rejected at the 0.10 level. There was no evidence of extinction effects. The final coordinates and temperature factors are listed in Tables I and 11. A final difference Fourier map showed residual electron densities which were centered around the cerium positions and ranged in value from -0.5 to 1.5 e/ A3. The largest peaks elsewhere were about 0.5 e/B3 with u ( p ) = 0.8 e/AB. None of these peaks was in a chemically reasonable position for a hydrogen atom. Table XI1 contains the observed and calculated structure factors.

+

Discussion The cerium atoms are in the speciaI positions (2c) surrounded by eight fluorine atoms in the conformation ( 8 ) W. C. Hamilton, A d a Cryst., 18, 502 (1965). (9) A. C. Larson, unpublished work. (10) H. T. Evans, A c f a Cryst., 14, 371 (1561). (11) D. T. Cromer and J. T. Waber, ibid., 18, 104 (1965). (12) D. T. Cromer, i b i d . , 18, 1 7 (1966).

Vol. 8, No. 1, January 1969

AMMONIUM HEXAFLUOROCERATE (IV) 35

TABLE I FRACTIONAL COORDINATES AND LEAST-SQUARES STANDARD ERRORS atom

Y

X

z

( 0)

.0414( 1)

.25

.0168(31)

.3430(10)

.4976(16)

F(1)

,2089(12)

.4816( 8)

.2019(12)

F(2)

. 1885(12)

. 1868(

7)

.2867(14)

F(3)

.9538(13)

.0979( 7)

. 5444(11)

Ce

.O

N

( 0)

t 323h IO' 0

I

a9 (20)

-10 (191 2 (10

Anisotropic temperature factors are defined by exp[ -(&h2 h k 2 P a l 2 Pnhk Pi,hZ pzrkl)].

+ +

+

+

around the Ce atom. The arrow Elipsoids are 3 times the rms ampli-

01

-12 (2:l)

a

Figure 1.-Coordination indicates the twofold axis. tude.

+

TABLE 111 OBSERVED A N D CALCULATED STRUCTURE FACTORS (COLUMN HEADINGS ARE k , 1, F,/K, [ Fol)

distances are given in Table IV. The average value of l / s for the F-F distance is 1.045, where s is the average distance between nearest atoms in the square of the antiprism and I is the average distance between nearest atoms in different squares.13 This value is very close to the value of 1.057 for a "most favorable" antiprism as calculated by Hoard and Silverton13 using a "neon shell" model for the ligands.

Ce - F(1)

- F(2) ce - P(3la c e - F(3) Ce

- h - F"(1) - - F'(2) F ' ( 1 ) - h - F(2)

of a slightly distorted square antiprism (see Figure 1). Each antiprism shares two fluorine atoms with each of the two neighboring antiprisms to form infinite chains in the direction of the c axis. Pertinent interatomic

TABLE 1lr INTERATOMIC DISTANCES IN (;"\;H&CeFe 2.198 (a) N - F' (1) 2.219 (8) N - F" (1) 2.298 (8) N - F' (2) 2 . 3 1 8 (8) N - F (2)

F'(1)

96.0 ( . 5 )

F'(1) N

85.0 ( . 5 )

2. 74 (2) 2 . 9 4 (2)

2 . 8 3 (2) 2 . 7 4 (2)

111,; (.j)

The nitrogen atom has four fluorine neighbors a t distances of 2.74-2.94 A plus five neighbors a t distances of 3.02-3.18 8. The six angles associated with the four shortest distances have an average value of 107.6", which suggests that the hydrogen atoms lie near these N-F vectors. It should also be noted that the fluorines bonded to only one cerium have the principal axis of their thermal ellipsoids tangent to the coordination polyhedron. For the shared fluorines the principal axis is perpendicular to the chaining direction. These are just the directions for which one would expect the largest vibrational amplitudes. The principal root-mean-square amplitudes of vibration for the atoms are shown in Table V. This structure is, in its over-all features, very simi(13) J. L. Hoard and J. V. Silverton, I?zovg. Chem., 2, 235 (1963).

36

ROBERT R. RYAN,ALLENC. LARSON, AND F. H. KRUSE

Inorganic Chemistry

I C

,-

,

Figure 2.-Stereoview

of the structure looking down the a axis, with the c direction toward the top of the picture. position.

Figure 3.-Stereoview

of the closely related Rb2UFe structure16 with the same orientation as Figure 2. indicates a uranium position.

TABLE IJ PRINCIPAL ROOT-MEAN-SQUARE AXPLITUDES Minimum

Intermediate

Maximum

.71

.89

N

1.05

1.72

2.20

F(1)

1.01

2.49

3.90

F(2)

1.14

2. 19

2.78

F(3)

.73

1.70

2.55

Ce

1.78

lar to that of both KzZrF614and RbzUFG,I6the principal difference being in the coordination polyhedra making up the chains. I n the latter two compounds the co-

C indicates the Ce

?IT indicates an Kb position and C

ordination polyhedra can best be described as somewhat distorted trigonal dodecahedra. Comparison of Figures 2 and 3 s h o w that distortion from what would be a trigonal dodecahedron in this structure to the existing antiprism is just that necessary t o make maximum use of H , . . F bonding of the fluorines to the ammonium ions. Acknowledgment.-Acknowledgment is due Dr. D. T. Cromer for his generous assistance and advice, as well as Dr. R. A. Penneman for the preparation of the compound. (14) H. Bode and G. Teufer, Acta Cvyst., 9, 929 (1956). (15) F. H. Kruse, to be submitted for publication.